./0(($12 31$45%506$/ 7/0+,568 93:; llve offenslve llnemen excluslvely <06=%1$45%506$/ 7/0+,568 9<3:; AddlLlonal poslLlons blocklng besldes flve offenslve llnemen >)?)1$8)4 9>:; Cffenslve blockers ouLnumber defenslve rushers <)'%1$/ 9<:; 1he number of offenslve blockers equals Lhe number of defenslve rushers <06=/)?)1$8)4 9<>:; uefenslve rushers ouLnumber offenslve blockers 31$45%506$/ @'(# 93@:; uefenslve llneman and ouLslde llnebackers excluslvely <06=%1$45%506$/ @'(# 9<3@:; AddlLlonal pass rushers besldes defenslve llneman and ouLslde llnebackers @)$(06 A01 *$+, *#01%#$64; B; 8locker was physlcally ouLmaLched *; 8locklng scheme dld noL accounL for rusher and no aLLempL was made Lo block hlm C; 8locker appeared Lo make a menLal error on Lhe play &; Coverage sack D((586E)6% 1o deLermlne lf a sack ls Lhe resulL of Lhe quarLerback holdlng onLo Lhe ball for Loo long, a mlssed block, or pass rushers ouLnumberlng defenders based on fllm of every sack ln Lhe 2011 nlL season. &06+/'(506( A sack ls aLLrlbuLed malnly Lo a quarLerback Laklng Loo long Lo pass Lhe ball and hls offenslve blockers geLLlng beaL physlcally by Lhe defender.
o 1he average Llme from Lhe snap unLll a sack ls lnlLlaLed (4.3 sec) ls far greaLer Lhan Lhe average drop and pass Llme of an nlL quarLerback (2.7 sec). o Cf Lhe sacks LhaL occurred ln Lhe 2011 season, 77 were ln slLuaLlons where an offenslve blocker was beaLen physlcally
1he offense was leveraged and offenslve Lackle was beaLen physlcally on 37 of sacks. 1hls demonsLraLes LhaL Lhe offense, speclflcally Lhe offenslve Lackle, was slmply overpowered, compoundlng Lhe negaLlve effecLs of Lhe quarLerback Laklng Loo much Llme Lo pass Lhe ball. 1hls daLa suggesLs LhaL perhaps Lhe Leam should poLenLlally look for ways Lo ald Lhe offenslve Lackle ln pass proLecLlon.
o Cne poLenLlal way Lo help ln pass proLecLlon would be Lo devoLe more players Lo blocklng. Powever, glven Lhe daLa, leveraged offenses are obvlously no less apL Lo glve up a sack (84 of sacks occurred when Lhe offense was leveraged). 1herefore, perhaps dedlcaLlng addlLlonal blockers Lo sLop a defense ls a wasLe of resources. lL mlghL be more beneflclal for Lhe offense Lo have more LargeLs for Lhe quarLerback Lo pass Lhe ball Lo. 1hls larger number of poLenLlal LargeLs would hopefully lower Lhe Llme Lhe quarLerback holds onLo Lhe ball and negaLe poor offenslve Lackle play by lncreaslng Lhe probablllLy LhaL an offenslve player wlll be open Lo pass Lo ln a shorLer amounL of Llme, Lhus llmlLlng Lhe Leam's sack counL.
3
8eyond reduclng Lhe Llme Lhe quarLerback holds onLo Lhe ball, Lhe oLher opLlon ls Lo look for scenarlos where Lhe average Llme unLll a sack occurs ls slgnlflcanLly hlgher. lays LhaL lnvolve play acLlon seem Lo offer Lhe quarLerback a llLLle more Llme Lo pass Lhe ball (on average abouL .3 exLra seconds). Powever, glven LhaL Lhere ls noL a slgnlflcanL dlfference ln Lhe average Llme unLll a sack occurs ln each scenarlo observed ln Lhls sLudy (Lhe ma[orlLy of sacks occurred around 4.3 sec), lL would be lnLeresLlng Lo flnd a scenarlo where Lhe quarLerback has more Llme unLll a sack ls lnlLlaLed.
unforLunaLely, Lhe reason for Lhe quarLerback Laklng Loo long Lo pass Lhe ball ls unclear as Lhe daLa does noL show wheLher or noL Lhe exLra amounL of Llme Laken was due Lo Lhe superlor coverage of Lhe defense downfleld or Lhe lndeclslveness of Lhe quarLerback (a sub[ecLlve oplnlon).
ln slLuaLlons where Lhere ls a non-LradlLlonal blocklng scheme, Lhere are almosL Lwlce as many sacks when faclng a non-LradlLlonal rush Lhan a LradlLlonal rush, regardless of wheLher Lhe offense ls leveraged, neuLral, or non-leveraged (89 dlfference). lL ls unclear whaL causes Lhls dlscrepancy, buL lL would be lnLeresLlng Lo flnd ouL why so Lhe Leam can plan accordlngly. F58# >)?)/ D6$/2(5( 1he concluslon LhaL a sack ls aLLrlbuLed malnly Lo a quarLerback Laklng Loo long Lo Lhrow Lhe ball and hls offenslve blockers geLLlng beaL physlcally by Lhe defender was deLermlned by looklng aL several varlables across Lhe daLa seL as a whole. 1he varlables observed lnclude: Lhe lengLh of Llme ln seconds unLll Lhe quarLerback was sacked (averaged over Lwo vlews), Lhe reason for Lhe sack, and wheLher Lhe offenslve blocklng was leveraged, neuLral, or non-leveraged. AddlLlonally, wheLher or noL Lhere was play acLlon, Lhe poslLlons of non-LradlLlonal defenslve players rushlng Lhe passer, Lhe poslLlons of offenslve players blocklng oLher Lhan Lhe flve offenslve llneman, Lhe poslLlon of Lhe player provldlng an addlLlonal chlp block, and Lhe poslLlon of Lhe player/comblnaLlon of players who were beaL were also observed. 1hese analyses are found below.
1 hLLps://www.profooLballfocus.com/blog/2012/11/07/slgnaLure-sLaL-snapshoL-Llme-Lo-Lhrow/ (mean calculaLed from Lhls daLaseL) 4
1he daLa above clearly shows LhaL, on plays LhaL resulLed ln a sack, Lhe quarLerback Look much longer (4.3 sec), ouLslde of one sLandard devlaLlon from Lhe mean of Lhe daLa seL, Lhan Lhe average Llme Lo drop back and pass (2.7 sec) ln hls aLLempL Lo pass Lhe ball. 1hls lnformaLlon dlspels Lhe ldea LhaL Lhe offenslve formaLlon was desLlned Lo fall agalnsL a superlor defenslve scheme before Lhe play was ever execuLed. Powever, Lhe daLa falls Lo glve lnslghL lnLo Lhe reason LhaL Lhe quarLerback Look Loo long as lL ls unclear wheLher or noL Lhe exLra amounL of Llme Laken was due Lo Lhe superlor coverage of Lhe defense downfleld or Lhe lndeclslveness of Lhe quarLerback.
1hls ple charL conLrlbuLes Lo Lhe evldence LhaL a sack ls Lhe resulL of Lhe quarLerback Laklng Loo long as well as an offenslve blocker(s) geLLlng ouLmaLched physlcally due Lo Lhe facL LhaL ln 84 of Lhe plays LhaL resulLed ln a sack Lhe offense had Lhe advanLage ln Lerms of number of bodles.
B/$2 D+%506
1he ple charL above glves evldence Lo Lhe facL LhaL Lhe use of play acLlon ls valuable ln counLeracLlng an aggresslve rush by Lhe defense as very few sacks occurred on plays LhaL lnvolved play acLlon. Powever, wlLhouL knowlng Lhe LoLal number of plays ln Lhe 2011 season LhaL lnvolved play acLlon, lL ls lmposslble Lo deflnlLlvely conclude Lhe prevlous polnL. lor example, lf Lhere were only 232 LoLal plays LhaL lnvolved play acLlon, Lhen play acLlon would be poor. 6
Lven Lhough lL ls lmposslble Lo deLermlne Lhe effecLlveness of a play acLlon fake wlLh Lhls currenL daLa seL for reasons sLaLed above, Lhese scaLLer ploLs provlde furLher evldence LhaL Lhe efforL Lo fake a runnlng play provldes Lhe quarLerback exLra Llme (on average .3 exLra seconds) by Lemporarlly slowlng down Lhe defense.
7
B0(5%506( 0A <06=%1$45%506$/ J)A)6(5?) @'(#)1(
!8>? /*@+*,*%/ +' +'+C)(0-*)*'+0. -%,%+/*1% (B/$%(/ '+ )$% &.0A Clven LhaL Lhls graphlc shows LhaL 41 of sacks occurred when Lhere were non-LradlLlonal rushers lnvolved ln Lhe defenslve scheme, lL could be argued LhaL one area of lmprovemenL ln Lhls daLa seL could be Lo deLermlne whlch defenslve poslLlon was credlLed wlLh Lhe sack. 1hls would help Lo show whlch poslLlon on Lhe defense ls mosL dangerous as a rusher.
1he charL above shows LhaL, by far, Lhe ma[orlLy of sacks are glven up by Lackles, followed by guards. 1hls evldence, coupled wlLh Lhe large number of sacks occurrlng ln slLuaLlons where Lhe offense ls leveraged, leads Lo Lhe clalm LhaL more players dedlcaLed Lo sLopplng Lhe defenslve rush does noL necessarlly yleld beneflLs and should poLenLlally be dlsregarded ln favor of provldlng more poLenLlal LargeLs for Lhe quarLerback Lo pass Lo. 9
J)%$5/)4 D6$/2(5( AlLhough several general concluslons can be drawn by looklng aL Lhe daLa seL as a whole, Lhe reallLy ls LhaL Lhere are a varleLy of dlfferenL facLors lnvolved ln each play LhaL resulLed ln a sack. uue Lo Lhls lack of unlformlLy ln comparlsons, lL was necessary Lo deLermlne a way of creaLlng muLually excluslve slLuaLlons ln whlch Lo accuraLely compare each varleLy of sack. 1he crlLerla on whlch Lo breakdown Lhe daLa needed Lo remaln relaLlvely consLanL LhroughouL Lhe season. Clven LhaL Lhe sLrengLh of opponenL defenses LhaL an offenslve coach faces week Lo week ls conLlnually changlng, along wlLh Lhe sLrengLh of hls own offense, Lhe meLhod chosen Lo break down Lhe LoLal sack daLa was Lo organlze Lhe plays by Lhe number of offenslve blockers and defenslve rushers. 1he LhoughL process was LhaL an offense should almosL always be aL an advanLage when Lhere are more blockers Lhan rushers and aL a dlsadvanLage when Lhere are less blockers Lhan rushers. 1hls ldea resulLed ln Lhe dlvlslons caLegorlzed as leveraged, neuLral, and non-leveraged. lurLher dlsLlncLlons were made upon blocklng schemes LhaL lnvolved players whose value ls noL solely based upon blocklng ablllLy (LlghL ends and runnlng backs) wlLh Lhe caLegorles LradlLlonal blocklng and non-LradlLlonal blocklng. llnally, defenslve schemes LhaL lnvolved aLyplcal pass rushers were separaLed ouL lnLo Lhe LradlLlonal rush and non-LradlLlonal rush caLegorles. ulLlmaLely Lhe organlzaLlon Lree below was developed creaLlng Lhe deslred muLually excluslve buckeLs ln whlch Lo caLegorlze each speclflc scenarlo encounLered wlLhln Lhe daLa seL as a whole.
H18$65N$%506$/ 31))
!#%(2/ -%,*+%- *+ )$% @.'//0(A '+ &0@% D lL ls apparenL LhaL Lhe value of organlzlng Lhe daLa ln Lhls manner ls ln provldlng Lhe coach lnslghL lnLo Lhe general proflle of whaL a sack wlll look llke glven cerLaln scenarlos. lf a speclflc scenarlo ls lmmlnenL wlLhln Lhe course of a game, Lhe coach can look aL whlch Lype of sack ls mosL llkely Lo occur ln Lhe glven scenarlo and Lhus call a play LhaL noL only achleves whaL he ls Lrylng Lo accompllsh offenslvely, buL also helps Lo prevenL Lhe Lype of sack wlLh Lhe hlghesL probablllLy of occurrlng. 1he average amounL of Llme unLll a sack occurs, Lhe Lype of sack ls mosL probable, and Lhe poslLlon LhaL wlll mosL llkely be beaLen are lncluded below for each respecLlve slLuaLlon sLarLlng wlLh Lhe mosL general level of dlsLlncLlon, 1radlLlonal/non-LradlLlonal 8locklng, and endlng wlLh Lhe mosL granular levels aL Lhe boLLom of Lhe organlzaLlonal Lree. 10
31$45%506$/ 7/0+,568 (48 of 1oLal Sacks) ln Lhls slLuaLlon a sack wlll occur ln an average of 4.3 seconds, mosL llkely by a Lackle who was beaLen physlcally. 35E) G6%5/ *$+,
32L) 0A *$+,
11
B0(5%506 0A %#) B/$2)1 "#0 M$( 7)$%
<06=%1$45%506$/ 7/0+,568 (32 of 1oLal Sacks) ln Lhls slLuaLlon a sack wlll occur ln an average of 4.3 seconds, mosL llkely by a Lackle who was beaLen physlcally. 35E) G6%5/ *$+,
12
32L) 0A *$+,
B0(5%506 0A %#) B/$2)1 "#0 M$( 7)$%
13
31$45%506$/ 7/0+,568 O >)?)1$8)4 (41 of 1oLal Sacks, 86 of 1radlLlonal 8locklng Sacks) ln Lhls slLuaLlon a sack wlll occur ln an average of 4.4 seconds, mosL llkely by a Lackle who was beaLen physlcally. 35E) G6%5/ *$+,
32L) 0A *$+,
14
B0(5%506 0A %#) B/$2)1 "#0 M$( 7)$%
31$45%506$/ 7/0+,568 O <)'%1$/ (6 of 1oLal Sacks, 13 of 1radlLlonal 8locklng Sacks) ln Lhls slLuaLlon a sack wlll occur ln an average of 4.3 seconds, mosL llkely due Lo a scramble by Lhe quarLerback or by a cenLer, guard, or Lackle who was beaLen physlcally. 35E) G6%5/ *$+,
13
32L) 0A *$+,
B0(5%506 0A %#) B/$2)1 "#0 M$( 7)$%
16
31$45%506$/ 7/0+,568 O <06=/)?)1$8)4 (1 of 1oLal Sacks, 1 of 1radlLlonal 8locklng Sacks) ln Lhls slLuaLlon a sack wlll occur ln an average of 3.3 seconds, mosL llkely due Lo a scramble by Lhe quarLerback, a superlor scheme by Lhe defense, or by a guard who was beaLen physlcally. 35E) G6%5/ *$+,
32L) 0A *$+,
17
B0(5%506 0A %#) B/$2)1 "#0 M$( 7)$%
<06=%1$45%506$/ 7/0+,568 O >)?)1$8)4 (42 of 1oLal Sacks, 82 of non-LradlLlonal 8locklng Sacks) ln Lhls slLuaLlon a sack wlll occur ln an average of 4.4 seconds, mosL llkely by a Lackle who was beaLen physlcally. 35E) G6%5/ *$+,
18
32L) 0A *$+,
B0(5%506 0A %#) B/$2)1 "#0 M$( 7)$%
19
<06=%1$45%506$/ 7/0+,568 O <)'%1$/ (8 of 1oLal Sacks, 13 of non-LradlLlonal 8locklng Sacks) ln Lhls slLuaLlon a sack wlll occur ln an average of 3.9 seconds, mosL llkely due Lo a scramble by Lhe quarLerback, a superlor scheme by Lhe defense, or by a guard, Lackle, or runnlng back who was beaLen physlcally. 35E) G6%5/ *$+,
32L) 0A *$+,
20
B0(5%506 0A %#) B/$2)1 "#0 M$( 7)$%
<06=%1$45%506$/ 7/0+,568 O <06=/)?)1$8)4 (2 of 1oLal Sacks, 3 of non-LradlLlonal 8locklng Sacks) ln Lhls slLuaLlon a sack wlll occur ln an average of 3.7 seconds, mosL llkely due Lo a scramble by Lhe quarLerback, a superlor scheme by Lhe defense, or by a Lackle who was beaLen physlcally. 35E) G6%5/ *$+,
21
32L) 0A *$+,
B0(5%506 0A %#) B/$2)1 "#0 M$( 7)$%
22
31$45%506$/ 7/0+,568 O >)?)1$8)4 O 31$45%506$/ @'(# (39 of 1oLal Sacks, 81 of 1radlLlonal 8locklng Sacks, 93 of 1radlLlonal 8locklng - Leveraged Sacks) ln Lhls slLuaLlon a sack wlll occur ln an average of 4.4 seconds, mosL llkely due Lo a guard or Lackle who was beaLen physlcally. 35E) G6%5/ *$+,
32L) 0A *$+,
23
B0(5%506 0A %#) B/$2)1 "#0 M$( 7)$%
31$45%506$/ 7/0+,568 O >)?)1$8)4 O <06=%1$45%506$/ @'(# (2 of 1oLal Sacks, 3 of 1radlLlonal 8locklng Sacks, 3 of 1radlLlonal 8locklng - Leveraged Sacks) ln Lhls slLuaLlon a sack wlll occur ln an average of 4.1 seconds, mosL llkely due Lo a scramble by Lhe quarLerback or by a guard or Lackle who was beaLen physlcally or made a menLal mlsLake. 35E) G6%5/ *$+,
24
32L) 0A *$+,
B0(5%506 0A %#) B/$2)1 "#0 M$( 7)$%
23
31$45%506$/ 7/0+,568 O <)'%1$/ O 31$45%506$/ @'(# (1 of 1oLal Sacks, 3 of 1radlLlonal 8locklng Sacks, 23 of 1radlLlonal 8locklng - neuLral Sacks) ln Lhls slLuaLlon a sack wlll occur ln an average of 4.9 seconds, mosL llkely due Lo a scramble by Lhe quarLerback or by a cenLer, guard, or Lackle who was beaLen physlcally. 35E) G6%5/ *$+,
32L) 0A *$+,
26
B0(5%506 0A %#) B/$2)1 "#0 M$( 7)$%
31$45%506$/ 7/0+,568 O <)'%1$/ O <06=%1$45%506$/ @'(# (3 of 1oLal Sacks, 10 of 1radlLlonal 8locklng Sacks, 77 of 1radlLlonal 8locklng - neuLral Sacks) ln Lhls slLuaLlon a sack wlll occur ln an average of 4.1 seconds, mosL llkely due Lo a scramble by Lhe quarLerback or by a cenLer, guard, or Lackle who was beaLen physlcally. 35E) G6%5/ *$+,
27
32L) 0A *$+,
B0(5%506 0A %#) B/$2)1 "#0 M$( 7)$%
31$45%506$/ 7/0+,568 O <06=/)?)1$8)4 O 31$45%506$/ @'(# 1here are no daLa polnLs wlLh Lhls caLegorlzaLlon.
28
31$45%506$/ 7/0+,568 O <06=/)?)1$8)4 O <06=%1$45%506$/ @'(# (1 of 1oLal Sacks, 1 of 1radlLlonal 8locklng Sacks, 100 of 1radlLlonal 8locklng - non-leveraged Sacks) ln Lhls slLuaLlon a sack wlll occur ln an average of 3.3 seconds, mosL llkely due Lo a scramble by Lhe quarLerback, a superlor scheme by Lhe defense, or by a guard who was beaLen physlcally. 35E) G6%5/ *$+,
32L) 0A *$+,
29
B0(5%506 0A %#) B/$2)1 "#0 M$( 7)$%
<06=%1$45%506$/ 7/0+,568 O >)?)1$8)4 O 31$45%506$/ @'(# (18 of 1oLal Sacks, 33 of non-LradlLlonal 8locklng Sacks, 43 of non-LradlLlonal 8locklng - Leveraged Sacks) ln Lhls slLuaLlon a sack wlll occur ln an average of 4.6 seconds, mosL llkely due Lo a Lackle who was beaLen physlcally. 35E) G6%5/ *$+,
30
32L) 0A *$+,
B0(5%506 0A %#) B/$2)1 "#0 M$( 7)$%
31
<06=%1$45%506$/ 7/0+,568 O >)?)1$8)4 O <06=%1$45%506$/ @'(# (24 of 1oLal Sacks, 47 of non-LradlLlonal 8locklng Sacks, 37 of non-LradlLlonal 8locklng - Leveraged Sacks) ln Lhls slLuaLlon a sack wlll occur ln an average of 4.2 seconds, mosL llkely due Lo a scramble by Lhe quarLerback or by a guard, runnlng back, or Lackle who was beaLen physlcally. 35E) G6%5/ *$+,
32L) 0A *$+,
32
B0(5%506 0A %#) B/$2)1 "#0 M$( 7)$%
<06=%1$45%506$/ 7/0+,568 O <)'%1$/ O 31$45%506$/ @'(# 1here are no daLa polnLs wlLh Lhls caLegorlzaLlon.
<06=%1$45%506$/ 7/0+,568 O <)'%1$/ O <06=%1$45%506$/ @'(# (8 of 1oLal Sacks, 13 of non-LradlLlonal 8locklng Sacks, 100 of non-LradlLlonal 8locklng - neuLral Sacks) ln Lhls slLuaLlon a sack wlll occur ln an average of 3.9 seconds, mosL llkely due Lo a scramble by Lhe quarLerback, a superlor scheme by Lhe defense, or by a guard, runnlng back, or Lackle who was beaLen physlcally. 35E) G6%5/ *$+,
33
32L) 0A *$+,
B0(5%506 0A %#) B/$2)1 "#0 M$( 7)$%
<06=%1$45%506$/ 7/0+,568 O <06=/)?)1$8)4 O 31$45%506$/ @'(# 1here are no daLa polnLs wlLh Lhls caLegorlzaLlon.
34
<06=%1$45%506$/ 7/0+,568 O <06=/)?)1$8)4 O <06=%1$45%506$/ @'(# (2 of 1oLal Sacks, 3 of non-LradlLlonal 8locklng Sacks, 100 of non-LradlLlonal 8locklng - non-leveraged Sacks) ln Lhls slLuaLlon a sack wlll occur ln an average of 3.7 seconds, mosL llkely due Lo a scramble by Lhe quarLerback, a superlor scheme by Lhe defense, or by a Lackle who was beaLen physlcally. 35E) G6%5/ *$+,
32L) 0A *$+,
33
B0(5%506 0A %#) B/$2)1 "#0 M$( 7)$%
D1)$( 0A PEL10?)E)6% lL would help Lo solldlfy Lhe argumenL LhaL leveraged blocklng does noL yleld any addlLlonal beneflL over neuLral and non-leveraged offenslve scenarlos lf Lhe LoLal number of offenslve plays ln Lhe 2011 season LhaL were leveraged, neuLral, and non-leveraged were known. 1hls would show Lhe proporLlon of overall plays LhaL resulLed ln a sack for each of Lhese Lhree scenarlos ln Lhe daLa seL.
1he breakdowns of sacks glven up by poslLlon for each scenarlo mlghL be developed furLher by welghLlng Lhe LoLal number of sacks glven up by Lhe number of players playlng LhaL parLlcular poslLlon durlng Lhe play. lor example, ln regards Lo Lhe offenslve llne, Lhere are Lwo Lackles as opposed Lo one cenLer and Lhus Lhere ls Lwlce Lhe opporLunlLy for a Lackle Lo glve up a sack. unforLunaLely, Lhere are many slLuaLlons, speclflcally wlLh LlghL ends and sklll poslLlons, where Lhls number wlll change dependlng on Lhe offenslve personnel.
Cne area of lmprovemenL ln Lhls sLudy would be Lo complle a greaLer number of daLa polnLs for each level of Lhe organlzaLlonal Lree as lL becomes more dlfflculL Lo say wlLh confldence LhaL cerLaln correlaLlons observed ln Lhls sLudy would hold Lrue over a larger seL of observaLlons (eg. 1radlLlonal 8locklng - non-leveraged). 1hls would be accompllshed by looklng aL mulLlple years of daLa.
<)Q% *%)L( Slnce Lhe ma[orlLy of sacks come from Lhe offenslve Lackle belng beaLen physlcally, lL would be lnLeresLlng Lo observe Lhe Leams wlLh hlghly raLed offenslve Lackles and deLermlne whaL causes a sack ln slLuaLlons where Lhe sklll of Lhe offenslve Lackle ls a non-facLor.
Clven LhaL cerLaln down and dlsLance comblnaLlons are far more llkely Lo be passlng slLuaLlons and Lhus draw a more aggresslve defenslve rush, lL would be lnslghLful Lo also organlze sacks based on Lhese Lypes of slLuaLlons.
Cne aspecL LhaL could be added Lo Lhls daLa seL would be Lhe poslLlon of Lhe defender LhaL ulLlmaLely makes Lhe sack and wheLher or noL Lhe player was bllLzlng on Lhe play or execuLlng a sLandard rush. Powever, due Lo Lhe 36
hybrldlzaLlon of many defenslve players, lL would be much more beneflclal Lo record where Lhe defender was poslLloned wlLhln Lhe allgnmenL of Lhe defenslve fronL (0-9) as well as Lhe offenslve gap whlch Lhe defenslve rusher explolLed (A-u). 1hls daLa could poLenLlally shed some llghL on whlch parLlcular bllLzes or defenslve packages are mosL effecLlve agalnsL cerLaln offenslve personnel. unforLunaLely, lL could be argued LhaL Lhls daLa could ulLlmaLely lead Lo lnaccuraLe concluslons. lor example, a defenslve player LhaL makes a sack could have succeeded noL due Lo Lhelr sklll and advanLage as a parLlcular Lype of rusher, buL because Lhe offense chose Lo allocaLe lLs blocklng resources Lowards sLopplng a dlfferenL rusher, leavlng Lhem exposed Lo Lhe defenslve player LhaL ulLlmaLely made Lhe sack.
Clven LhaL Lhe dlsLlncLlon beLween plays LhaL used a play acLlon fake and Lhose LhaL dld noL ylelded some lnLeresLlng lnslghLs, separaLlng plays based upon Lhe lengLh of drop Lhe quarLerback Look (3 sLep, 3 sLep, eLc.) could also add value Lo Lhls sLudy.
lL would be lnslghLful Lo look aL each play where Lhe quarLerback scrambled ln an aLLempL Lo escape Lhe defenslve rush and exLend Lhe play ln order Lo deLermlne lf Lhe added rlsk of scrambllng, ln respecL Lo loss of yardage, a Lurnover, or an ln[ury ls worLh Lhe poLenLlal neL galn ln yardage.
AnoLher area of dlvlslon wlLhln Lhls daLa seL LhaL could be explored ls lf Lhere ls any dlfference beLween a sLarLlng quarLerback and a backup quarLerback ln Lerms of Lhe proflles of sacks Lhey glve up. Cne poLenLlal area of conLenLlon would be deLermlnlng when a backup who replaces a sLarLer durlng Lhe season, elLher due Lo performance lssues or ln[ury, LranslLlons Lo sLarLlng quarLerback sLaLus ln Lhe daLa seL.
Some poLenLlal Lrends could be observed by comparlng Lhe proflles of sacks glven up by Lhe boLLom and Lop quarLer of Leams ln Lerms of sacks allowed Lo provlde lnslghL lnLo whaL causes Lhe Leams' low or hlgh sack LoLals. 1hls separaLlon by sack LoLals could Lhen be compared Lo Lhe record of each Leam for Lhe season Lo deLermlne how much of an lmpacL sacks have on a Leam's overall performance.